Aim/PurposeThis work intends to give a method for the automatic superimposition of facial and cranium anatomical images coupled with integrating jaw movement. Using an automated alignment method will help to raise the accuracy and efficiency of the forensic face reconstruction procedure. Given their reliance on human participation, conventional approaches are prone to subjectivity and errors. Differential Genetic Algorithm (DGA) accounts for mandibular articulation and allows for exact alignment of skull and facial images, therefore reaching strong optimization.

BackgroundForensic face reconstruction is a crucial field of research for the anthropological sciences and the criminal justice system. Although modern methods offer benefits, their dependability is not always guaranteed since they rely on human interaction. By using a DGA, the proposed approach overcomes this limit and boosts efficiency. Differential evolution and genetic algorithms, which can capture all the special features required for perfect face reconstruction, help to improve the alignment.

MethodologyThis study aims to enhance the alignment parameters between image graphs of the skull and visage, and it also considers mandibular articulation using a DGA. Genetic operators and differential evolution support the program in efficiently investigating the domain of feasible solutions. Whether the superimposed images properly depict the intended face traits is found rather successfully by means of the fitness function.

ContributionThis work offers a suitable solution for progressive forensic facial reconstruction using a technique based on DGA for automated overlay. An improved level of accuracy and realism is shown by comparing the obtained result with other existing approaches and methods on mandibular articulation in the reconstructed facial images.

FindingsThe proposed DGA has been proven to match images of the face and the cranium exactly by including the articulation of the jaw. The automatic overlay shows the possibilities of the forensic techniques since it generates results equal to or better than those acquired by hand.

Recommendation for ResearchersScholars should improve the proposed method by means of more dataset integration and genetic algorithm configuration change.

Future ResearchIn future research, this work can be enhanced using several deep learning algorithms to achieve better accuracy and performance.